Writing method based on big data and smart pen

Abstract

This invention provides a writing method and smart pen based on big data-assisted writing. It receives test data from a writing user, acquires the characters to be trained from the test data, retrieves a training database to perform similarity analysis on the characters to be trained, and obtains the predicted training counts from the deviation state to the standard state for each character to be trained. It then acquires the writing user's historical training data, and obtains adjustment data based on the predicted training counts and actual training counts for each character to be trained in the historical training data. Based on the adjustment data, it adjusts the predicted training counts to obtain the planned training counts for the corresponding characters to be trained. Based on the planned training counts and the baseline writing counts, it obtains a customized writing page corresponding to the writing user. Finally, it customizes the combinations of the characters to be trained in the customized writing page according to a reconstructed lexicon, generating a combined writing page corresponding to the writing user.

Description

Technical Field

[0001] This invention relates to data processing technology, and more particularly to a writing method for assisting writing based on big data and a smart pen. Background Technology

[0002] In order to improve the standard of their handwriting, people often use auxiliary tools to practice writing and improve the standard of their handwriting.

[0003] Currently, people generally buy paper calligraphy practice books for writing practice. However, the number of characters available for practice in paper versions is limited, and it is impossible to judge the standard of the characters written by the writer. Therefore, it is impossible to flexibly change the number of characters available for practice according to the actual situation of the user. For example, some characters are already written standardly enough, but there are also characters with very poor standard, which require a lot of practice. However, the number of characters available for practice in paper versions is fixed, which will affect the standardization and efficiency of writing practice.

[0004] Therefore, how to customize the corresponding writing page based on the user's own handwriting style and improve the efficiency of practicing writing standards has become an urgent problem to be solved. Summary of the Invention

[0005] This invention provides a writing method and a smart pen based on big data-assisted writing. It can judge the standard of the handwriting based on the user's own handwriting and customize the number of practice characters, thereby improving the efficiency of practicing writing standards.

[0006] A first aspect of the present invention provides a writing method based on big data-assisted writing, comprising:

[0007] Receive test data from writing users, obtain the characters to be trained in the test data, retrieve the training database to perform similarity analysis on the characters to be trained, and obtain the predicted number of training times from the deviation state to the standard state corresponding to each character to be trained.

[0008] Obtain the historical training data of the writing user, and obtain adjustment data based on the predicted training times and actual training times of each character to be trained in the historical training data;

[0009] Based on the adjustment data, the predicted training times are adjusted to obtain the planned training times for the corresponding character to be trained. Based on the planned training times and the baseline writing times, a customized writing page corresponding to the writing user is obtained.

[0010] Based on the reconstructed lexicon, the characters to be trained in the customized writing page are customized and combined to generate a combined writing page corresponding to the writing user.

[0011] Optionally, in one possible implementation of the first aspect, receiving test data from the writing user and obtaining the characters to be trained from the test data includes:

[0012] Obtain the test data of the writing user, and retrieve the standard writing data corresponding to each test character in the test data;

[0013] Based on the comparison information between the actual writing data of the test character and the corresponding standard writing data, the deviation writing data is obtained. The number of deviation writing data is counted to obtain the deviation writing quantity, and the number of the corresponding actual writing data is counted to obtain the actual writing quantity.

[0014] The deviation ratio is obtained based on the ratio of the number of deviations written to the number of actual writings. When the deviation ratio is greater than a preset ratio, the corresponding test character is used as the character to be trained.

[0015] Optionally, in one possible implementation of the first aspect, obtaining the deviation writing data based on the comparison information between the actual written data of the test character and the corresponding standard written data includes:

[0016] Obtain the writing pixel value corresponding to the actual writing data of the test character, and statistically analyze the dot matrix coordinates corresponding to the actual writing data based on the writing pixel value to obtain the writing coordinate set corresponding to the actual writing data.

[0017] Retrieve the standard coordinate set corresponding to the corresponding standard writing data, determine the intersection of the standard coordinate set and the corresponding writing coordinate set, and obtain the standard coordinate set;

[0018] Based on the difference between the writing coordinate set and the standard coordinate set, the deviation coordinate set is obtained. The number of lattice coordinates in the deviation coordinate set is obtained to obtain the number of writing deviations. The number of lattice coordinates in the standard coordinate set is obtained to obtain the number of writing standards.

[0019] The writing deviation ratio is obtained by comparing the number of writing deviations with the number of writing standards.

[0020] When the writing deviation ratio is determined to be greater than the preset writing ratio, the corresponding actual writing data is used as the deviation writing data.

[0021] Optionally, in one possible implementation of the first aspect, the step of retrieving the training database to perform similarity analysis on the characters to be trained, and obtaining the predicted training times from the deviation state to the standard state corresponding to each of the characters to be trained, includes:

[0022] Retrieve from the training database the writing training users corresponding to the character to be trained from the deviation state to the standard state, and the number of writing training times corresponding to the writing training users;

[0023] Stroke analysis is performed on the actual writing data to obtain the current deviation strokes. Based on the current deviation strokes, similarity screening is performed on the writing training users to obtain writing similar users corresponding to each intersecting stroke.

[0024] The number of writing training times corresponding to the writing type of users is determined as the number of writing type. Based on the number of writing type of each intersecting stroke and the corresponding initial weight value, the predicted number of training times corresponding to the character to be trained is obtained.

[0025] Optionally, in one possible implementation of the first aspect, the step of performing stroke analysis on the corresponding actual writing data to obtain the current deviation stroke includes:

[0026] Obtain the set of stroke coordinates corresponding to the actual writing data, and retrieve the standard set of strokes corresponding to the standard writing data, as well as the number of dot matrix coordinates in the standard set of strokes to obtain the number of standard dots.

[0027] The stroke specification set is obtained by the intersection of the stroke standard set and the stroke coordinate set;

[0028] Based on the difference between the set of stroke coordinates and the corresponding set of stroke specifications, a set of stroke deviations is obtained, and the number of dot matrix coordinates in the set of stroke deviations is obtained to obtain the number of deviation dots.

[0029] The number of stroke errors corresponding to the corresponding writing stroke is calculated based on the standard number of dots, the preset number of dots, and the preset number of errors.

[0030] When the number of deviation dots is determined to be greater than the number of strokes tolerable, the corresponding written stroke is regarded as a misaligned stroke.

[0031] The number of misaligned strokes is determined, and the stroke deviation ratio corresponding to the misaligned strokes is obtained based on the ratio of the number of misaligned strokes to the corresponding actual number of written strokes.

[0032] When the stroke deviation ratio is determined to be greater than the preset stroke ratio, the corresponding misaligned stroke is taken as the deviation stroke.

[0033] Optionally, in one possible implementation of the first aspect, the step of performing similarity filtering on the writing training users based on the current deviation strokes to obtain writing users of the same type corresponding to each current deviation stroke count includes:

[0034] Obtain the deviation strokes of the writing training users as historical deviation strokes, and determine the corresponding writing training users as initial screening training users based on the intersection relationship between the historical deviation strokes and the current deviation strokes.

[0035] The strokes that intersect with the current deviation strokes of the initial screening training users are obtained as intersecting strokes, and the number of intersecting strokes is obtained as the number of deviation strokes. The initial screening training users are sorted in descending order according to the number of deviation strokes to obtain the selection user sequence.

[0036] Based on a preset selection quantity, the initial screening training users in the selected user sequence are selected sequentially to obtain the writing type users corresponding to each of the intersecting strokes.

[0037] Optionally, in one possible implementation of the first aspect, obtaining the historical training data of the writing user and obtaining adjustment data based on the predicted training times and actual training times of each character to be trained in the historical training data includes:

[0038] Obtain the predicted training times and actual training times for each character to be trained from the historical training data of the writing user;

[0039] When it is determined that the actual number of training iterations is greater than the predicted number of training iterations, a positive training value is determined based on the difference between the actual number of training iterations and the predicted number of training iterations.

[0040] The initial weight value corresponding to the largest number of deviation strokes is obtained as the positive training weight. The positive training weight is increased and adjusted according to the positive training value to obtain the adjusted positive training weight.

[0041] The weight adjustment value is obtained based on the difference between the adjusted positive training weight and the positive training weight.

[0042] The initial weight values ​​are sorted in ascending order based on the number of deviation strokes to obtain a weight adjustment sequence;

[0043] The first initial weight value in the weight adjustment sequence is selected as the reverse training weight, and the adjusted reverse training weight is obtained based on the difference between the reverse training weight and the weight adjustment value.

[0044] When it is determined that the adjusted reverse training weight is less than 0, the first initial weight value in the weight adjustment sequence is deleted, the repeated adjustment value is obtained according to the absolute value of the adjusted reverse training weight, and the corresponding adjusted reverse training weight is set to 0.

[0045] Use the repeated adjustment value as the current weight adjustment value, and repeat the above steps of using the repeated adjustment value as the current weight adjustment value until the adjusted reverse training weight is greater than or equal to 0, thus obtaining multiple adjusted reverse training weights.

[0046] Adjusted data is obtained based on the adjusted forward training weights and the adjusted backward training weights.

[0047] Optionally, in one possible implementation of the first aspect, obtaining the historical training data of the writing user and obtaining adjustment data based on the predicted training times and actual training times of each character to be trained in the historical training data includes:

[0048] When it is determined that the actual number of training iterations is less than the predicted number of training iterations, a reverse training value is determined based on the difference between the predicted number of training iterations and the actual number of training iterations.

[0049] The initial weight value corresponding to the largest number of deviation strokes is obtained as the reverse training weight. The reverse training weight is adjusted by decreasing the reverse training value to obtain the adjusted reverse training weight.

[0050] The weight adjustment value is obtained based on the difference between the reverse training weight and the adjusted reverse training weight;

[0051] The initial weight value corresponding to the minimum number of deviation strokes is obtained as the positive training weight. The adjusted positive training weight is obtained based on the weight adjustment value and the sum of the positive training weight.

[0052] Adjusted data is obtained based on the adjusted forward training weights and the adjusted backward training weights.

[0053] Optionally, in one possible implementation of the first aspect, obtaining the customized writing page corresponding to the writing user based on the planned training count and the baseline writing count includes:

[0054] The total number of training iterations corresponding to the characters to be trained is obtained by counting the planned training iterations. Based on the ratio of each planned training iteration to the total number of training iterations, the proportion coefficient corresponding to the characters to be trained is obtained.

[0055] The customized number of writings corresponding to the character to be trained is obtained by multiplying the proportion coefficient and the baseline number of writings.

[0056] The corresponding characters to be trained are combined according to the number of customized writing attempts to obtain a customized writing page.

[0057] Optionally, in one possible implementation of the first aspect, the step of customizing the combination of characters to be trained in the customized writing page according to the reconstructed lexicon to generate a combined writing page corresponding to the writing user includes:

[0058] Based on the preset word groups in the reconstructed lexicon, the characters to be trained in the customized writing page are combined to obtain multiple customized word groups, each of which includes multiple characters to be trained.

[0059] The number of characters to be trained in the customized word group is obtained to get the number of combinations. The customized word group is sorted in descending order according to the number of combinations to get the customized sequence.

[0060] Select the first customized word group in the customized sequence as the current customized word group, and take the character to be trained in the current customized word group as the current character to be trained.

[0061] Obtain the customized writing count corresponding to the current character to be trained, select the current character to be trained with the smallest customized writing count as the deletion and update character, and use the remaining current characters to be trained as relay update characters;

[0062] The number of times the customized writing of the deleted and updated characters is used as the update count. The current customized word group is generated based on the update count, and the number of times the customized writing of the relay updated characters is updated based on the update count.

[0063] Delete the custom word groups that contain the deleted and updated characters in the custom sequence, and repeat the above steps of deleting the custom word groups that contain the deleted and updated characters in the custom sequence until the custom sequence is empty, to obtain multiple current custom word groups, and use the remaining characters to be trained as random characters to be trained;

[0064] The combined writing page is obtained by combining the currently customized phrases and the random characters to be trained.

[0065] Optionally, in one possible implementation of the first aspect, it also includes:

[0066] Obtain the number of initial screening users for initial screening. When the number of initial screening users is equal to 0, determine the basic stroke unit corresponding to the current deviation stroke. Extract the corresponding writing strokes in the characters to be trained as selection strokes based on the basic stroke unit. Delete and update the selection strokes based on the current deviation strokes to obtain strokes of the same type.

[0067] Obtain the maximum, minimum, maximum, and minimum values ​​of the horizontal coordinates, vertical coordinates, and ordinates of the corresponding dot matrix coordinates for each stroke of the same type.

[0068] The horizontal length value of the same type of stroke is obtained by the difference between the maximum and minimum horizontal coordinate values ​​of the same type of stroke; the vertical length value of the same type of stroke is obtained by the difference between the maximum and minimum vertical coordinate values ​​of the same type of stroke.

[0069] Obtain the maximum, minimum, maximum, and minimum values ​​of the horizontal and vertical coordinates of the corresponding dot matrix coordinates for each of the current deviation strokes;

[0070] The current horizontal length value is obtained based on the difference between the maximum and minimum horizontal coordinate values ​​of the current deviation stroke, and the current vertical length value is obtained based on the difference between the maximum and minimum vertical coordinate values ​​of the current deviation stroke.

[0071] A first approximate evaluation value is obtained based on the absolute value of the difference between the current horizontal length value and the horizontal length values ​​of the same type;

[0072] A second approximate evaluation value is obtained based on the absolute value of the difference between the current longitudinal length value and the longitudinal length value of the same type;

[0073] The evaluation coefficient corresponding to the same type of stroke is calculated based on the first approximate evaluation value and the second approximate evaluation value.

[0074] Based on the comparison information between the evaluation coefficient and the preset evaluation interval, the corresponding strokes of the same type are determined as approximate strokes;

[0075] Based on the approximate strokes, the writing training users are selected to obtain writing users of the same type as the intersecting strokes.

[0076] A second aspect of the present invention provides a smart writing pen for assisted writing based on big data, comprising:

[0077] The analysis module is used to receive test data from the writing user, obtain the characters to be trained in the test data, retrieve the training database to perform similarity analysis on the characters to be trained, and obtain the predicted number of training times from the deviation state to the standard state corresponding to each character to be trained.

[0078] The acquisition module is used to acquire the historical training data of the writing user and obtain adjustment data based on the predicted training times and actual training times of each character to be trained in the historical training data.

[0079] The adjustment module is used to adjust the predicted training times based on the adjustment data to obtain the planned training times for the corresponding character to be trained, and to obtain a customized writing page corresponding to the writing user based on the planned training times and the baseline writing times.

[0080] The generation module is used to customize and combine the characters to be trained in the customized writing page according to the reconstructed lexicon, and generate a combined writing page corresponding to the writing user.

[0081] A third aspect of the present invention provides an electronic device comprising: a memory, a processor, and a computer program, the computer program being stored in the memory, and the processor executing the computer program to perform the methods described in the first aspect of the present invention and various possible methods related to the first aspect.

[0082] A fourth aspect of the present invention provides a storage medium storing a computer program, which, when executed by a processor, is used to implement the first aspect of the present invention and various methods possibly involved in the first aspect.

[0083] The beneficial effects of this invention are as follows:

[0084] 1. This invention can judge the standard of handwriting based on the individual's own handwriting style and customize the number of practice characters to improve the efficiency of handwriting standard practice. At the same time, this invention can also customize the characters to be trained and display them in the form of phrases to customize and update the customized writing page, so that the writing user can increase the memory of phrases during the writing training process.

[0085] 2. This invention can obtain the corresponding characters to be trained from the test data of writing users and perform similarity analysis to obtain the predicted training times for each character. This allows for customized writing training to be generated for writing users, thereby improving the efficiency of writing training. Specifically, this invention can compare the actual writing data of writing users with standard writing data to identify the characters that need to be trained. Furthermore, this invention can determine the stroke deviation of the characters to be trained, allowing for the identification of the corresponding writing user based on the current stroke deviation, thus obtaining the relevant writing training times. This ensures a more accurate prediction of the training times for the characters to be trained. Finally, this invention can also statistically analyze the historical writing training data of the writing user and adjust the predicted training times accordingly, making the determined customized writing times more accurate, thereby improving the writing training efficiency of the writing user.

[0086] 3. This invention can combine characters to be trained based on a vocabulary database, displaying them as phrases on the combined writing page. This allows users to strengthen their memory of phrases during writing training. Specifically, this invention can sort customized phrases according to the number of combinations, obtaining a customized sequence. This allows for prioritizing the selection of phrases with the most combinations for subsequent page customization, enhancing phrase memory and improving training efficiency. Furthermore, this invention can use the current character to be trained corresponding to the minimum number of customized writing attempts as the deletion and update character, and the remaining current characters to be trained as relay update characters. This ensures that characters meeting the customized training attempts are deleted, preventing duplicate updates on the page and avoiding exceeding the customized training attempts. Attached Figure Description

[0087] Figure 1 A flowchart of a writing method based on big data-assisted writing provided by the present invention;

[0088] Figure 2 A schematic diagram of a standard writing data provided by the present invention;

[0089] Figure 3 A schematic diagram of deviation writing data provided by the present invention;

[0090] Figure 4 A schematic diagram of the structure of a smart writing pen based on big data-assisted writing provided by the present invention;

[0091] Figure 5 This is a schematic diagram of the hardware structure of an electronic device provided by the present invention. Detailed Implementation

[0092] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0093] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0094] This invention provides a writing method based on big data-assisted writing, such as... Figure 1 As shown, it includes:

[0095] S1. Receive the test data of the writing user, obtain the characters to be trained in the test data, retrieve the training database to perform similarity analysis on the characters to be trained, and obtain the predicted training times from the deviation state to the standard state corresponding to each character to be trained.

[0096] It can be understood that the writing personnel can preferably write a paragraph of text, so that the standard degree of the text can be judged according to the content written by the personnel, and then the corresponding practice times can be calculated for non-standard fonts, so as to achieve efficient writing practice.

[0097] Among them, the test data is the text content written by the personnel themselves, the characters to be trained are the characters that need to be practiced, the training database is the database that statistically records the writing situations of other users, which can be constructed by the server statistically recording the writing situations of each user, the deviation state is the state where the written font is non-standard, the standard state is the standard state of the written font, and the predicted training times are the predicted number of times of practice required to reach the standard state.

[0098] Through the above implementation manner, the present invention can obtain the predicted training times of the characters to be trained, so as to facilitate subsequent customized training.

[0099] In some embodiments, in step S1 (receiving the test data of the writing user, obtaining the characters to be trained in the test data, retrieving the training database to perform similarity analysis on the characters to be trained, and obtaining the predicted training times from the deviation state to the standard state corresponding to each character to be trained), it specifically includes:

[0100] S11. Obtain the test data of the writing user, and retrieve the standard writing data corresponding to each test character in the test data.

[0101] It can be understood that the test data of the writing user is obtained, and thus the standard writing data corresponding to each test character in the test data is retrieved, so as to facilitate subsequent comparison of the standard writing data with the data actually written by the user, thereby determining the characters that need to be practiced, facilitating targeted training, and improving the efficiency of writing practice.

[0102] Among them, the test characters are each character in the test data, and the standard writing data is the standard written text corresponding to each character. For example, as Figure 2 shown, it is the standard writing data of the test character "ten thousand" written by the writing user in the test data.

[0103] S12. According to the comparison information between the actual writing data of the test character and the corresponding standard writing data, obtain the deviation writing data, count the number of the deviation writing data to obtain the deviation writing quantity, and count the number of the corresponding actual writing data to obtain the actual writing quantity.

[0104] It can be understood that by comparing the text actually written by the user with the standard writing data, text with non-standard font writing can be obtained. For example, as Figure 3 shown in the deviation writing data "ten thousand".

[0105] Among them, the actual writing data is the text data actually written by the writing user, the deviation writing data is the non-standard text data of the font writing, the deviation writing quantity is the quantity of the non-standard text of the written font, and the actual writing quantity is the quantity of the text in the actual data.

[0106] For example: when the test data written by the writing user is 1 line of "ten thousand" with a total of 10 characters, among which, the actual writing quantity corresponding to "ten thousand" in the actual writing data is 10. When compared with the standard writing data, it is judged that there are 5 corresponding to the non-standard character "ten thousand", then the deviation writing quantity can be obtained as 5.

[0107] In some embodiments, in step S12 (obtaining the deviation writing data according to the comparison information between the actual writing data of the test character and the corresponding standard writing data), it specifically includes:

[0108] S121, obtaining the writing pixel values corresponding to the actual writing data of the test character, and based on the writing pixel values, counting the dot matrix coordinates corresponding to the actual writing data to obtain a writing coordinate set corresponding to the actual writing data.

[0109] It should be noted that when the writing user performs text writing, text writing can be carried out on special paper with a dot matrix pen. The dot matrix pen uses infrared recognition technology, utilizes the paper with dot matrix codes and infrared high-speed camera capture technology, and records the dot matrix coordinates on the paper during the writing process, so as to synchronize the information paper screen after writing on the paper, so that the written information can be synchronized in the electronic device. Therefore, a dot matrix coordinate set of each character in the actual writing process of the user can be obtained. For example, the writing font is black, and the standard font is red or dotted line, etc. This is the prior art and will not be elaborated here.

[0110] It can be understood that when text writing is carried out, corresponding writing traces will be left. Therefore, according to the difference between the writing pixel values and the pixel values of the cardboard without text writing, the coordinates of the text written by the writing user can be obtained, so as to obtain the writing coordinate set of the actual writing data, so as to compare with the coordinate set of the standard writing data subsequently, and thus judge the standardness of the actual writing data of the writing user.

[0111] Among them, the writing pixel value is the text pixel value corresponding to the actual writing data, the dot matrix coordinate is the coordinate of the dot matrix code corresponding to the written text, and the writing coordinate set is the set of multiple dot matrix coordinates corresponding to the actual writing data.

[0112] S122, retrieve the standard coordinate set corresponding to the corresponding standard writing data, determine the intersection of the standard coordinate set and the corresponding writing coordinate set, and obtain the specification coordinate set.

[0113] It can be understood that the standard coordinate set is the coordinate set corresponding to the standard written text font. For example, the set of coordinates corresponding to the standard font "ten thousand". The specification coordinate set is the intersection of the coordinate set of the actually written font and the coordinate set of the standard font. It is not difficult to understand that the part where the intersection occurs is the part with correct writing.

[0114] Through the above implementation manner, the present invention can obtain the corresponding specification coordinate set, so as to facilitate subsequent standardization judgment of the actually written text.

[0115] S123, according to the difference set between the writing coordinate set and the specification coordinate set, obtain the deviation coordinate set, obtain the number of dot matrix coordinates in the deviation coordinate set to get the writing deviation quantity, and obtain the number of dot matrix coordinates in the standard coordinate set to get the writing standard quantity.

[0116] It can be understood that the deviation coordinate set is the difference set between the writing coordinate set and the specification coordinate set, which is the difference set between the coordinate sets of the text actually written by the writing user and the standard text, so that the deviation degree between the actually written text and the standard written text can be determined according to the deviation coordinate set.

[0117] Among them, the writing deviation quantity is the number of dot matrix coordinates in the deviation coordinate set, and the writing standard quantity is the number of dot matrix coordinates in the standard coordinate set.

[0118] Through the above implementation manner, the present invention can obtain the writing deviation quantity and the writing standard quantity, so that subsequently, according to the writing deviation quantity and the writing standard quantity, the deviation degree corresponding to the actual writing data can be judged, and thus the text that needs to be trained can be customized for the writing user.

[0119] S124, according to the ratio of the writing deviation quantity and the writing standard quantity, obtain the writing deviation ratio.

[0120] It can be understood that the writing deviation ratio is the ratio of the writing deviation quantity and the writing standard quantity, so that subsequently, the standard degree of the actual writing data can be determined according to the writing deviation ratio.

[0121] S125. When it is determined that the writing deviation ratio is greater than the preset writing ratio, take the corresponding actual writing data as the deviation writing data.

[0122] It can be understood that the preset writing ratio is a preset ratio value, which can be set artificially in advance, such as 1 / 5.

[0123] For example: when obtaining the coordinates of the character "wan" written by the writing user, and the corresponding writing deviation ratio is 1 / 3, and the preset writing ratio is 1 / 5, then the actual writing data "wan" can be taken as the deviation writing data.

[0124] S13. According to the ratio of the deviation writing quantity to the actual writing quantity, obtain the deviation ratio. When it is determined that the deviation ratio is greater than the preset ratio, take the corresponding test character as the character to be trained.

[0125] It can be understood that the deviation ratio is the ratio of the deviation writing quantity to the actual writing quantity, and the preset ratio is a preset ratio value, which can be set artificially, for example, 1 / 10. It can be explained that occasional non-standard characters will not be judged as deviation characters. When the deviation ratio is greater than the preset ratio, it can be explained that the writing user does not write the corresponding characters standardly and needs to strengthen training.

[0126] For example, when the preset ratio is 1 / 10, the deviation writing quantity is 5, and the actual writing quantity is 10, the corresponding deviation ratio obtained is 1 / 2. Since the deviation ratio is greater than the preset ratio, it can be explained that half of the "wan" characters in these 10 "wan" characters are non-standard. Therefore, it can be determined that the corresponding "wan" character can be used as the character to be trained, and the writing user needs to be trained to make the user's writing more standard.

[0127] Through the above implementation manner, the present invention can obtain the deviation writing data to facilitate the customized determination of the training content for the writing user.

[0128] S I4. Retrieve the writing training user corresponding to the character to be trained in the training database from the deviation state to the standard state, and the writing training times corresponding to the writing training user.

[0129] It can be understood that when the writing training user conducts writing practice, the relevant training data of the user's writing from the deviation state to the standard state will be saved in the training database. Thus, when the character to be trained corresponding to the writing user is the same, the corresponding writing training user in the training database can be retrieved, and the corresponding writing training times can be obtained.

[0130] Among them, the writing training user is the user who has completed the writing training saved in the training database, and the number of writing training times is the number of times the writing training user practices the characters to be trained. For example, when the character to be trained is 'ten thousand', retrieve the writing training user A who has also practiced the character 'ten thousand', and obtain that the number of writing training times corresponding to the practice from the deviation state to the standard state is 100.

[0131] Through the above implementation manners, the present invention can obtain the number of writing training times corresponding to the writing training user, so as to make the customized training plan for the writing user more accurate.

[0132] S15. Perform stroke analysis on the corresponding actual writing data to obtain the current deviation strokes, and perform similarity screening on the writing training users according to the current deviation strokes to obtain the writing similar users corresponding to each cross stroke.

[0133] It can be understood that in the actual characters written by the writing user, maybe a certain stroke in the character is relatively non-standard. Therefore, the strokes can be analyzed, so as to more accurately determine the problem of non-standard writing, and similarity screening can be performed on the writing training users according to the strokes, and users whose deviation from the actual writing data of the writing user is more similar can be selected, so that the subsequent determined number of training times can be more accurate.

[0134] Among them, the current deviation strokes are the strokes with non-standard writing of characters. For example, as Figure 3 shown, when the first stroke 'one horizontal line' and the second stroke 'a left-falling stroke' of the character 'ten thousand' written by the writing user deviate from the standard strokes, the current deviation strokes are the first and second strokes, and subsequently, the users corresponding to the strokes with cross-overlap can be analyzed.

[0135] In some embodiments, step S15 (performing stroke analysis on the corresponding actual writing data to obtain the current deviation strokes, and performing similarity screening on the writing training users according to the current deviation strokes to obtain the writing similar users corresponding to each cross stroke) specifically includes:

[0136] S1501. Obtain the set of stroke coordinate points corresponding to the writing strokes of the actual writing data, retrieve the set of stroke standards corresponding to the corresponding standard writing data, and obtain the standard dot matrix quantity of the dot matrix coordinates in the set of stroke standards.

[0137] It can be understood that the set of stroke coordinate points is the set of dot matrix coordinates of the written strokes, the set of stroke standards is the set of coordinates of the writing strokes corresponding to the standard writing data, and the standard dot matrix quantity is the quantity of dot matrix coordinates in the set of stroke standards.

[0138] Through the above embodiments, the present invention can obtain a set of stroke coordinates and a set of stroke standards, which are convenient for subsequent processing, so as to judge the strokes of the actual writing data of the user.

[0139] S1502. Obtain a set of stroke specifications according to the intersection of the set of stroke standards and the set of stroke coordinates.

[0140] It can be understood that the set of stroke standards and the set of stroke coordinates are intersected to obtain the corresponding intersection as the set of stroke specifications, so as to determine the part of the stroke standards in the written text, and then the part with writing deviation in the corresponding stroke can also be obtained subsequently.

[0141] S1503. Based on the difference set between the set of stroke coordinates and the corresponding set of stroke specifications, obtain a set of stroke deviations, obtain the number of dot matrix coordinates in the set of stroke deviations, and obtain the number of deviation dot matrices.

[0142] It can be understood that the set of stroke coordinates and the corresponding set of stroke specifications are compared to obtain the difference set as the set of stroke deviations, and the number of dot matrix coordinates in the set of stroke deviations is counted to obtain the number of deviation dot matrices, so as to judge the standard degree of the corresponding stroke.

[0143] Among them, the number of deviation dot matrices is the number of dot matrix coordinates in the set of stroke deviations.

[0144] S1504. Calculate according to the number of standard dot matrices, the preset number of dot matrices and the preset error tolerance number, and obtain the number of stroke error tolerances corresponding to the corresponding writing stroke.

[0145] It should be noted that for a smaller stroke, such as the dot in the character "太", if there is a deviation in the position, it will cause the overall font to be less neat, and the allowable error is smaller. However, for a larger stroke, such as "一", the user is allowed to have errors such as their own writing stroke. Therefore, the present invention will determine the corresponding error tolerance number according to the size of the standard stroke.

[0146] It can be understood that the preset number of dot matrices is a number set artificially according to the actual situation, and the preset error tolerance number is a number corresponding to the preset number of dot matrices, which is set artificially in advance. For example, when the preset number of dot matrices is 10, the corresponding preset error tolerance number can be 1, which means that for the dot matrix coordinates in the stroke, 1 error is allowed for every 10 dot matrix coordinates.

[0147] Among them, the number of stroke error tolerances is the number of dot matrices allowed to have errors in the stroke.

[0148] For example: when the standard dot matrix quantity is 100, the preset dot matrix quantity is 10, and the preset fault tolerance quantity is 1, the stroke fault tolerance quantity can be obtained as (100 / 10)×1 = 10. For the convenience of understanding, only an example is given here.

[0149] It is not difficult to understand that when there are only very few dot matrix coordinates in the strokes of the actual writing data written by the user that are inconsistent with the standard dot matrix coordinates, it does not affect the standardization of the strokes. Therefore, the corresponding stroke fault tolerance quantity can be calculated through the standard dot matrix quantity, the preset dot matrix quantity, and the preset fault tolerance quantity, so as to facilitate subsequent standard judgment of the strokes.

[0150] Specifically, the calculation method is to obtain a ratio coefficient according to the ratio of the standard dot matrix quantity and the preset dot matrix quantity, and according to the ratio coefficient, obtain the stroke fault tolerance quantity corresponding to the corresponding writing stroke through the product of the ratio coefficient and the preset fault tolerance quantity. Among them, the preset dot matrix quantity has a corresponding preset fault tolerance quantity.

[0151] S1505, when it is determined that the deviation dot matrix quantity is greater than the stroke fault tolerance quantity, the corresponding writing stroke is taken as a misaligned stroke.

[0152] It can be understood that when the deviation dot matrix quantity of the corresponding stroke is greater than the stroke fault tolerance quantity, it can be explained that the deviation degree of the stroke is outside the allowable range, so the writing stroke can be taken as a misaligned stroke.

[0153] Among them, an incorrect stroke is a stroke with a non-standard degree of writing.

[0154] For example: when the deviation dot matrix quantity of the writing stroke "丿" is 40 and the stroke fault tolerance quantity is 10, the deviation dot matrix quantity 40 > the stroke fault tolerance quantity 10, then the writing stroke "丿" can be taken as a misaligned stroke.

[0155] S1506, determine the misalignment quantity of the misaligned stroke, and obtain the stroke deviation ratio corresponding to the misaligned stroke according to the ratio of the misalignment quantity and the corresponding actual writing quantity.

[0156] It can be understood that when there are 10 "ten thousand" characters in the actual writing data, and the second stroke "丿" is a misaligned stroke, the number of errors in the second stroke will be counted among these 10 "ten thousand" characters. When 5 out of 10 have errors in the second stroke "丿", the corresponding misalignment quantity is 5. The stroke deviation ratio is the ratio of the misalignment quantity corresponding to the incorrect stroke and the actual writing quantity. For example, when the misalignment quantity of the misaligned stroke "丿" is 5 and the actual writing quantity is 10, the corresponding stroke deviation ratio can be obtained as 1 / 2, so as to further judge the standardization of the font of the corresponding actual writing data. When it is judged that the deviation degree of the font is relatively large, the customized training plan can be made more accurate.

[0157] S1507, when it is determined that the stroke deviation ratio is greater than the preset stroke ratio, the corresponding misaligned stroke is taken as the deviation stroke.

[0158] It can be understood that the preset stroke ratio is a pre-set ratio value. When the stroke deviation ratio is greater than the preset stroke ratio, the corresponding misaligned stroke is taken as the deviation stroke.

[0159] S1508, obtain the deviation strokes of the writing training user as historical deviation strokes, and determine the corresponding writing training user as a preliminary screening training user according to the cross relationship between the historical deviation strokes and the current deviation strokes.

[0160] It can be understood that the historical deviation strokes are the deviation strokes of the writing training user, and the preliminary screening training user is a writing training user with deviations in the strokes of the same character.

[0161] For example: when the historical deviation strokes of writing training user A in the character "万" are the first and third strokes "一" and "丿", the historical deviation strokes of writing training user B in the character "万" are "一" and "", the historical deviation strokes of writing training user C in the character "万" are "" and "丿", the historical deviation strokes of writing training user D in the character "万" are "", and the current deviation strokes are "一" and "丿", then according to the cross relationship, that is, the relationship that the current deviation strokes exist in the historical deviation strokes, it can be obtained that the writing training users with the deviation strokes "一" and "丿" are writing training users A, B, and C, so as to determine that the preliminary screening training users are writing training users A, B, and C.

[0162] S1509, obtain the strokes that cross between the preliminary screening training user and the current deviation strokes as the cross strokes, and obtain the number of the cross strokes as the number of deviation strokes. Sort the preliminary screening training users in descending order according to the number of deviation strokes to obtain the selected user sequence.

[0163] It can be understood that the cross strokes are the strokes where the historical deviation strokes and the current deviation strokes in the initial screening training users cross, and the number of deviation strokes is the number of crosses between the historical deviation strokes and the current deviation strokes in the initial screening training users. For example, when the historical deviation strokes of writing training user A in the character "万" are "一" and "丿", the historical deviation strokes of writing training user B in the character "万" are "一" and "", and the historical deviation strokes of writing training user C in the character "万" are "" and "丿", where the current deviation strokes are the first stroke "一" and the third stroke "丿", then it can be obtained that the corresponding number of deviation strokes of initial screening training user A is 2, the corresponding number of deviation strokes of initial screening training user B is 1, and the corresponding number of deviation strokes of initial screening training user C is 1. The cross strokes can be the first stroke, the second stroke, or the first and second strokes.

[0164] Among them, the selected user sequence is the sequence obtained by sorting the initial screening training users in descending order according to the number of deviation strokes. For example, the selected user sequence can be (initial screening training user A, initial screening training user B, initial screening training user C).

[0165] Through the above implementation manners, the present invention can obtain the corresponding selected user sequence, so as to subsequently determine the number of training times closer to the training situation of the current writing user, making the obtained number of training times more accurate.

[0166] S1510, successively select the initial screening training users in the selected user sequence based on a preset selection quantity, and obtain the writing similar users corresponding to each of the cross strokes.

[0167] It can be understood that the preset selection quantity is the selection quantity preset by a person, which can be 3. Thus, when there are many users with similar situations, it can avoid and select a certain number of users according to the actual situation for data acquisition and processing, and can reduce the amount of data processing.

[0168] Among them, the writing similar users are the users after selecting the initial screening training users.

[0169] For example, when the number of deviation strokes is 2, the corresponding writing similar user is initial screening training user A. When the number of deviation strokes is 1, the corresponding writing similar users are initial screening training user B and initial screening training user C. For the convenience of understanding, only examples are given for the number of times.

[0170] Through the above implementation manners, the present invention can obtain the writing similar users, so as to subsequently determine the corresponding number of training times.

[0171] S16. Determine the number of writing training times corresponding to the writing similar users as the writing similar times, and perform calculation processing based on the writing similar times and the corresponding initial weight values of each of the cross strokes, to obtain the predicted training times corresponding to the to-be-trained characters.

[0172] It can be understood that the writing similar times are the number of writing training times corresponding to the writing similar users. For example, when the writing similar user A writes 100 times and achieves the transition from the deviation state to the standard state, the writing similar times corresponding to this user is 100.

[0173] Among them, the initial weight value is the weight value preset artificially. It is not difficult to understand that the initial weight values corresponding to different numbers of deviation strokes can be different. When the number of deviation strokes of the writing similar user is larger, it indicates that the deviation situation between the writing similar user and the writing user is more similar, and the corresponding initial weight value can be larger. When the number of deviation strokes of the writing similar user is smaller, it indicates that the similarity of the deviation situation between the writing similar user and the writing user is not high, and the corresponding initial weight value can be smaller. For example, when the deviation strokes of the writing user for the character "wan" are the first stroke "one" and the third stroke "丿", and the historical deviation strokes of the writing similar user A are also the first stroke "one" and the third stroke "丿", then the deviation situations of the two are similar, and the writing training times corresponding to the writing similar user A are more reference-worthy, thus making the training times customized for the writing user more accurate.

[0174] It is not difficult to understand that the predicted training times are the number of times predicted through calculation that the writing user needs to write to write the to-be-trained character from the deviation state to the standard state.

[0175] The predicted training times are obtained through the following formula.

[0176]

[0177] Among them, is the predicted training times, is the type number of the cross strokes, is the number of writing similar users corresponding to the th cross stroke, is the writing similar times of the th writing similar user under the th cross stroke,

[0178] is the number of types of cross strokes. There are 3 types of cross strokes. For example, in the character "wan", the first stroke and the second stroke, the first stroke, the second stroke.

[0179] It is understandable that the present invention will average the number of training times of each user corresponding to the types of cross strokes and multiply by their respective weights to obtain the predicted training times. That is, the average number of training times of each user who selects the first stroke and the second stroke is calculated and multiplied by the corresponding weight; the average number of training times of each user who selects the first stroke is calculated and multiplied by the corresponding weight; the average number of training times of each user who selects the second stroke is calculated and multiplied by the corresponding weight. The sum of these three is used to obtain the corresponding training times, and the sum of their respective weights is equal to 1.

[0180] S2. Obtain the historical training data of the writing user, and obtain adjustment data based on the predicted training times and actual training times of each to-be-trained character in the historical training data.

[0181] It can be understood that the writing user can perform writing training on other characters in the past. Therefore, based on the predicted training times and actual training times of each character in the historical training data of the writing user, the predicted training times of the current to-be-trained character can be adjusted, and then a more accurate training time can be obtained, which can save the training time of the writing user and realize the writing training of the writing user.

[0182] Among them, the historical training data is the data of the writing user's writing practice in the past, and the actual training times are the number of times the writing user actually trains the to-be-trained character to reach the standard state through writing training.

[0183] For example: when the predicted number of times for the to-be-trained character in the historical training data of the writing user is 300, but the writing user actually writes 200 times and completes the writing training, the predicted training times of the current to-be-trained character can be adjusted based on the predicted training times and actual training times of the historical to-be-trained character, so as to obtain a more accurate training time.

[0184] In some embodiments, step S2 (obtaining the historical training data of the writing user and obtaining adjustment data based on the predicted training times and actual training times of each to-be-trained character in the historical training data) specifically includes:

[0185] A21. Obtain the predicted training times and actual training times of each to-be-trained character in the historical training data of the writing user.

[0186] Understandably, it is possible to prioritize obtaining the predicted and actual training counts for each character to be trained from the historical training data of the writing user, in order to perform data calculations and obtain adjusted data, making the training counts of the writing user more accurate.

[0187] A22, when it is determined that the actual number of training iterations is greater than the predicted number of training iterations, a positive training value is determined based on the difference between the actual number of training iterations and the predicted number of training iterations.

[0188] It is understandable that when users practice writing the characters to be trained, the actual number of practice sessions may be more than the predicted number of training sessions in order to complete the writing training of the characters to be trained. When the actual number of training sessions is greater than the predicted number of training sessions, a positive training value can be obtained, which is convenient for subsequent data adjustment. It is worth mentioning that, generally speaking, the more incorrect strokes there are, the greater the corresponding number of training sessions.

[0189] Among them, the positive training value is the value that increases the number of training iterations in a positive direction.

[0190] For example, if the actual number of training iterations is 300 and the predicted number of training iterations is 200, then the positive training value is 300-200=100.

[0191] A23, obtain the initial weight value corresponding to the largest number of deviation strokes as the positive training weight, and adjust the positive training weight according to the positive training value to obtain the adjusted positive training weight.

[0192] It is understandable that the larger the number of deviated strokes, the closer it is to the actual writing situation of the user, and the greater the reference value of the number of training times determined. Therefore, the initial weight value corresponding to the largest number of deviated strokes can be used as the positive training weight.

[0193] For example, if the positive training weight is 60%, and the positive training value is large, the positive training weight can be increased to obtain an adjusted positive training weight of 80%.

[0194] It's easy to understand that when the positive training value is larger, it means that the number of training iterations needs to be increased significantly. Correspondingly, the positive training weights will be increased based on the positive training value, resulting in adjusted positive training weights, which makes the subsequent determined number of training iterations more accurate.

[0195] The adjusted positive training weights are obtained using the following formula.

[0196]

[0197] in, These are the adjusted positive training weights. For positive training weights, This refers to the actual number of training sessions. To predict the number of training iterations, To adjust the value, it's easy to understand that the actual number of training iterations... The larger the value, the more appropriate the adjusted positive training weights. The larger the value, the more strokes are involved, and the more training sessions are required. This results in a greater overall number of training sessions, and the adjustment value will vary. These weights are pre-set by the user, allowing for adaptive training to obtain weight values ​​that are tailored to the user.

[0198] Since the overall weight sum is 1, the weight of subsequent strokes with fewer strokes will be adaptively reduced.

[0199] A24. The weight adjustment value is obtained based on the difference between the adjusted positive training weight and the positive training weight.

[0200] Understandably, the weight adjustment value is the difference between the adjusted positive training weight and the original positive training weight. For example, if the adjusted positive training weight is 80% and the original positive training weight is 60%, then the corresponding weight adjustment value is 80% - 60% = 20%, which facilitates subsequent weight adjustments and allows for the calculation of the adjustment data.

[0201] A25, based on the number of deviation strokes, sort the initial weight values ​​in ascending order to obtain the weight adjustment sequence.

[0202] It is understandable that the weight adjustment sequence is a sequence obtained by sorting the initial weight values ​​in ascending order according to the number of deviation strokes. For example, it can be (20%, 20%, 60%).

[0203] A26, select the first initial weight value in the weight adjustment sequence as the reverse training weight, and obtain the adjusted reverse training weight based on the difference between the reverse training weight and the weight adjustment value.

[0204] Understandably, once the initial weight value corresponding to the largest number of deviated strokes is adjusted, the remaining initial weight values ​​need to be adjusted accordingly, so that multiple initial weight values ​​in the weight adjustment sequence can be adjusted sequentially.

[0205] The reverse training weight is the first initial weight value in the weight adjustment sequence. For example, when the weight adjustment sequence is (20%, 20%, 60%), the corresponding reverse training weight is 20%. The adjusted reverse training weight is the difference between the reverse training weight and the weight adjustment value. For example, when the reverse training weight is 20% and the weight adjustment value is 20%, the corresponding adjusted reverse training weight is 20% - 20% = 0.

[0206] A27, when it is determined that the adjusted reverse training weight is less than 0, delete the first initial weight value in the weight adjustment sequence, obtain the repeated adjustment value according to the absolute value of the adjusted reverse training weight, and set the corresponding adjusted reverse training weight to 0.

[0207] It is understandable that when the weight adjustment sequence is (10%, 30%, 60%), the corresponding reverse training weight is 10%, and when the weight adjustment value is 20%, the corresponding adjusted reverse training weight is 10%-20%=-10%<0. Therefore, the first initial weight value of 10% in the weight adjustment sequence is deleted, so that the repeated adjustment value can be obtained, and the corresponding adjusted reverse training weight is set to 0.

[0208] The repeated adjustment value is the absolute value of the adjusted reverse training weight. For example, when the adjusted reverse training weight is 10%-20%=-10%, the repeated adjustment value is 10%.

[0209] Through the above implementation method, the present invention obtains repeated adjustment values, which facilitates the subsequent adjustment of the next initial weight value in the weight adjustment sequence.

[0210] A28, take the repeated adjustment value as the current weight adjustment value, repeat the above steps of taking the repeated adjustment value as the current weight adjustment value until the adjusted reverse training weight is greater than or equal to 0, and obtain multiple adjusted reverse training weights.

[0211] Understandably, when the repeated adjustment value is 10%, 10% can be used as the current weight adjustment value so that the next initial weight value in the weight adjustment sequence can be adjusted. For example, after the above adjustment, the current weight adjustment sequence is (0, 30%, 60%), so the second initial weight of 30% is adjusted, and the adjusted reverse training weight is 30%-10%=20%>0, so the adjustment of the weight value in the weight adjustment sequence is stopped. Similarly, when the weight adjustment sequence is (20%, 20%, 60%), after adjusting the first initial weight, the adjusted reverse training weight is 0, so the adjustment of other weight values ​​is stopped.

[0212] A29. Adjusted data is obtained based on the adjusted forward training weights and the adjusted backward training weights.

[0213] It is understandable that the adjusted data refers to the adjusted forward training weights and the adjusted backward training weights. For example, if the weights before adjustment are 60%, 30%, and 10%, then the weights after adjustment can be 80%, 20%, and 0.

[0214] Through the above implementation methods, the present invention can obtain corresponding adjustment data, thereby facilitating subsequent adjustments to the predicted training count based on the adjustment data, and thus obtaining a more accurate training count.

[0215] In some embodiments, step S2 (obtaining the historical training data of the writing user, and obtaining adjustment data based on the predicted training times and actual training times of each character to be trained in the historical training data) specifically includes:

[0216] B21, when it is determined that the actual number of training iterations is less than the predicted number of training iterations, the reverse training value is determined based on the difference between the predicted number of training iterations and the actual number of training iterations.

[0217] It is understandable that when users practice writing the characters to be trained, the actual number of practice sessions may be less than the predicted number of training sessions. By completing the writing training of the characters to be trained with a small number of practice sessions, a reverse training value can be obtained when the actual number of training sessions is less than the predicted number of training sessions, which is convenient for subsequent data adjustment.

[0218] The reverse training value is the value that reduces the number of training iterations.

[0219] For example, if the actual number of training iterations is 150 and the predicted number of training iterations is 200, then the reverse training value is 200-150=50.

[0220] B22, obtain the initial weight value corresponding to the largest number of deviation strokes as the reverse training weight, and adjust the reverse training weight by reducing it according to the reverse training value to obtain the adjusted reverse training weight.

[0221] It is understandable that the larger the number of deviated strokes, the closer it is to the actual writing situation of the user, and the greater the reference value of the number of training times determined. Therefore, the initial weight value corresponding to the largest number of deviated strokes can be used as the reverse training weight.

[0222] For example, when the back training weight is 60%, since the back training value is the difference between the predicted number of training times and the actual number of training times, the back training weight can be reduced to obtain an adjusted back training weight of 40%.

[0223] It's easy to understand that when the back training value is larger, it means that the number of training iterations needs to be adjusted downwards significantly. Correspondingly, the back training weights will be reduced based on the back training value to obtain the adjusted back training weights, thereby making the subsequent determined number of training iterations more accurate.

[0224] The adjusted inverse training weights are obtained using the following formula.

[0225]

[0226] in, The adjusted reverse training weights, For reverse training of weights, This refers to the actual number of training sessions. To predict the number of training iterations, To adjust the value, it's easy to understand that the actual number of training iterations... The smaller the value, the more appropriate the adjusted reverse training weights. The smaller the value, the fewer training sessions are needed for strokes with a large number of strokes, resulting in fewer overall training sessions, and the adjustment value varies. These weights are pre-set by the user, allowing for adaptive training to obtain weight values ​​that are tailored to the user.

[0227] B23. The weight adjustment value is obtained based on the difference between the reverse training weight and the adjusted reverse training weight.

[0228] Understandably, the weight adjustment value is the difference between the adjusted reverse training weight and the adjusted reverse training weight. For example, if the adjusted reverse training weight is 40% and the reverse training weight is 60%, then the corresponding weight adjustment value is 60% - 40% = 20%, which facilitates subsequent weight adjustments and allows for the calculation of the adjustment data.

[0229] B24. Obtain the initial weight value corresponding to the minimum number of deviation strokes as the positive training weight. Based on the weight adjustment value and the sum of the positive training weights, obtain the adjusted positive training weights.

[0230] It is understandable that the positive training weight is the initial weight value corresponding to the minimum number of deviation strokes, and the adjusted positive training weight is the sum of the weight adjustment value and the positive training weight. For example, when the positive training weight is 20% and the weight adjustment value is 20%, the corresponding adjusted positive training weight is 40%.

[0231] B25, based on the adjusted forward training weights and the adjusted backward training weights, the adjusted data is obtained.

[0232] Through the above embodiments, the present invention can obtain corresponding adjustment data, so as to facilitate subsequent adjustment of the predicted training times according to the adjustment data, and further obtain more accurate training times.

[0233] S3. Adjust the predicted training times based on the adjustment data to obtain the planned training times corresponding to the characters to be trained. According to the planned training times and the reference writing times, obtain a customized writing page corresponding to the writing user.

[0234] It can be understood that the predicted training times are adjusted according to the adjustment data to obtain relatively accurate planned training times for the characters to be trained. Furthermore, each character to be trained can be used to customize the writing practice page based on the corresponding planned training times and the reference writing times per page, so as to obtain a customized writing page.

[0235] Among them, the planned training times are the adjusted times of the predicted training times, and the reference writing times are the times of character writing training that can be carried out per page.

[0236] Through the above embodiments, the present invention can obtain a corresponding customized writing page, so as to facilitate subsequent targeted training for the writing user and improve the writing training efficiency of the writing user.

[0237] In some embodiments, step S3 (obtaining a customized writing page corresponding to the writing user according to the planned training times and the reference writing times) specifically includes:

[0238] S31. Count the planned training times corresponding to the characters to be trained to obtain the total training times. According to the ratio of each planned training time to the total training times, obtain the proportion coefficient corresponding to the characters to be trained.

[0239] It can be understood that the total training times are the sum of the planned training times corresponding to each character to be trained. For example, when the planned training times for the character "ten thousand" to be trained are 150, the planned training times for the character "li" to be trained are 100, and the planned training times for the character "thousand" to be trained are 50, the total training times can be obtained as 150 + 100 + 50 = 300.

[0240] Among them, the proportion coefficient is the ratio of the planned training times to the total training times. For example, when the planned training times for the character "wan" to be trained is 150, the planned training times for the character "li" to be trained is 100, the planned training times for the character "qian" to be trained is 50, and the total training times is 300, then the proportion coefficient for the character "wan" to be trained can be obtained as 150 / 300 = 1 / 2, the proportion coefficient for the character "li" to be trained is 100 / 300 = 1 / 3, and the proportion coefficient for the character "qian" to be trained is 50 / 300 = 1 / 6.

[0241] Through the above implementation manners, the present invention can obtain the proportion coefficients corresponding to each character to be trained, so as to perform customized planning on the page for the characters to be trained.

[0242] S32. Obtain the customized writing times corresponding to the character to be trained according to the product of the proportion coefficient and the reference writing times.

[0243] It can be understood that the number of characters available for training per page is limited, such as 300 characters. Therefore, in order to enable the user to perform writing training on different characters to be trained on each page, the proportion coefficients corresponding to each character to be trained can be multiplied by the reference writing times per page, and the times of each character to be trained on each page can be obtained.

[0244] Among them, the customized writing times is the product of the proportion coefficient and the reference writing times.

[0245] For example: when the proportion coefficient of the character "wan" to be trained is 1 / 2, the proportion coefficient of the character "li" to be trained is 1 / 3, the proportion coefficient of the character "qian" to be trained is 1 / 6, and the reference writing times is 300, then the customized writing times of the character "wan" to be trained can be obtained as 1 / 2×300 = 150, the customized writing times of the character "li" to be trained is 1 / 3×300 = 100, and the customized writing times of the character "qian" to be trained is 1 / 6×300 = 50. It can be shown that the number of times the character "wan" can be written by the writing user on each page is 150, the number of times the character "li" can be written by the writing user is 100, and the number of times the character "qian" can be written by the writing user is 50.

[0246] S33. Combine the corresponding characters to be trained according to the customized writing times to obtain a customized writing page.

[0247] It can be understood that the customized writing page is a page of characters to be trained with customized writing times.

[0248] Through the above implementation manners, the present invention can obtain a customized writing page for the writing user to perform writing training.

[0249] S4. Based on the reconstructed lexicon, the characters to be trained in the customized writing page are customized and combined to generate a combined writing page corresponding to the writing user.

[0250] Understandably, reconstructing the lexicon into a database with word groups can be pre-configured and used to combine characters to be trained, thereby displaying them in the form of word groups. This can help writing users strengthen their memory of word groups during the writing training process.

[0251] Among them, the combined writing page is the page obtained by combining the characters to be trained into words.

[0252] In some embodiments, step S4 (customizing and combining the characters to be trained in the customized writing page according to the reconstructed lexicon to generate a combined writing page corresponding to the writing user) specifically includes:

[0253] S41, combine the characters to be trained in the customized writing page according to the preset word groups in the reconstructed lexicon to obtain multiple customized word groups, the customized word groups including multiple characters to be trained.

[0254] Understandably, preset phrases are pre-set phrases, such as "ten thousand miles", "ten million miles", "mileage", etc., while customized phrases are phrases composed of characters to be trained. For example, if the characters to be trained are "ten thousand", "mile", and "thousand", then customized phrases can be obtained as "ten million miles", "ten thousand miles", "ten million miles", etc.

[0255] S42, obtain the number of characters to be trained in the customized word group, obtain the number of combinations, sort the customized word group in descending order according to the number of combinations, and obtain the customized sequence.

[0256] Understandably, the number of combinations refers to the number of characters to be trained in the customized phrase. For example, when the customized phrase is "ten thousand miles", the corresponding number of combinations is 3, and when the customized phrase is "ten thousand miles", the corresponding number of combinations is 2. When the characters to be trained are combined, multiple phrases can be obtained. In order to ensure that each phrase contains as many characters to be trained as possible and improve the training efficiency of the writing user, the customized phrases can be sorted according to the number of combinations to obtain a customized sequence. This allows the phrases with the most combinations to be selected for page customization in the future.

[0257] The customized sequence is a sequence of words obtained by sorting the customized words in descending order according to the number of combinations. For example, the customized sequence can be (ten thousand miles, ten thousand miles, ten thousand miles).

[0258] It's easy to understand that when the number of combinations is the same, custom word groups with the same number of combinations can be randomly arranged.

[0259] S43. Select the first customized phrase in the customized sequence as the current customized phrase, and use the characters to be trained in the current customized phrase as the current characters to be trained.

[0260] It can be understood that the current customized phrase is the first customized phrase in the customized sequence. For example, when the customized sequence is (ten thousand li, ten thousand li, ten million), "ten thousand li" can be selected as the current customized phrase. Among them, the current characters to be trained are the characters to be trained in the current customized phrase. Furthermore, the current characters to be trained in the current customized phrase are "thousand", "ten thousand", and "li".

[0261] Through the above implementation, the present invention can obtain the corresponding current characters to be trained, which is convenient for subsequent customized update of the page.

[0262] S44. Obtain the corresponding customized writing times of the current characters to be trained, select the current character to be trained corresponding to the smallest customized writing times as the deletion and update character, and use the remaining current characters to be trained as the relay update characters.

[0263] It can be understood that in order to facilitate subsequent deletion of the characters to be trained that meet the customized training times, and to avoid repeated updates on the page, resulting in the characters to be trained updated on the page exceeding the customized training times. Therefore, the current character to be trained corresponding to the smallest customized writing times can be used as the deletion and update character, and the remaining current characters to be trained can be used as the relay update characters.

[0264] For example: when the corresponding customized writing times of the obtained current characters to be trained "thousand", "ten thousand", and "li" are 50, 150, and 100 respectively, the current character to be trained "thousand" corresponding to the customized writing times of 50 can be used as the deletion and update character, and "ten thousand" and "li" can be used as the relay update characters.

[0265] S45. Use the customized writing times of the deletion and update characters as the update times, generate the current customized phrase according to the update times, and update the customized writing times of the relay update characters based on the update times.

[0266] It can be understood that the update times are the customized writing times of the deletion and update characters. For example, when the customized writing times of the deletion and update characters are 50, the corresponding update times are 50. Thus, 50 current customized phrases "ten thousand li" can be generated. Since the generated "ten thousand li" already has 50 relay update characters "ten thousand" and "li", the customized writing times of the relay update characters "ten thousand" and "li" can be updated. For example, the updated customized writing times of "ten thousand" is 150 - 50 = 100, and the updated customized writing times of "li" is 100 - 50 = 50.

[0267] S46. Delete the custom phrases in the custom sequence that contain the deletion update character, and repeat the above step of deleting the custom phrases in the custom sequence that contain the deletion update character until the custom sequence is empty, obtaining multiple current custom phrases, and use the remaining characters to be trained as random characters to be trained.

[0268] It can be understood that since the deletion update characters corresponding to the custom writing times have been generated, there is no need to continue selecting custom phrases with deletion update characters. Therefore, the custom phrases in the custom sequence that contain the deletion update character can be deleted. For example, when 50 custom writing phrases of "ten thousand miles" have been generated, the update of 50 characters to be trained for the character "thousand" is completed. Thus, there is no need to continue selecting the character "thousand" for update and training on the page. Therefore, the "ten thousand miles" and "ten thousand" in the custom sequence (ten thousand miles, ten thousand, thousand miles) can be deleted.

[0269] It is not difficult to understand that when the remaining characters to be trained have not reached the custom writing times, the step of repeatedly selecting custom phrases until deleting the custom phrases in the custom sequence that contain the deletion update character can be repeated until the custom sequence is empty, obtaining multiple current custom phrases. When the custom sequence is empty and there are still characters to be trained that have not been updated to the corresponding times, the remaining characters to be trained can be used as random characters to be trained, so as to be separately generated on the page for the writing user to practice writing.

[0270] For example: when the custom sequence is (ten thousand miles), "ten thousand miles" can be used as the current custom phrase, and 50 "ten thousand miles" are generated to update the character to be trained "miles". Then the corresponding custom writing times for the character to be trained "ten thousand" become 100 - 50 = 50. Therefore, "ten thousand" can be used as a random character to be trained.

[0271] S47. Combine the current custom phrases and the random characters to be trained to obtain a combined writing page.

[0272] It can be understood that the combined writing page is a page generated by combining the current custom phrases and random characters to be trained. For example, 300 characters can be generated on the page for writing training. Then 50 "ten thousand miles", 50 "ten thousand miles", and 50 "ten thousand" are generated and combined on the page to obtain a combined writing page for the writing user to practice writing.

[0273] It should be noted that when selecting writing training users, there may be no writing training users that meet the deviation strokes. Therefore, in order to obtain relevant data for reference and make the determined custom writing times more accurate and in line with the writing training of writing users, the following is also included:

[0274] C01. Obtain the initial screening quantity of the users for initial screening training. When it is determined that the initial screening quantity is equal to 0, determine the basic stroke unit corresponding to the current deviation stroke. According to the basic stroke unit, extract the corresponding writing strokes in the character to be trained as the selected strokes. Delete and update the selected strokes according to the current deviation stroke to obtain similar strokes.

[0275] It can be understood that the basic stroke unit is the basic unit stroke, for example, horizontal, vertical, left-falling, right-falling. Similar strokes are similar strokes of the same type, for example, the first and second strokes in "待" are both left-falling strokes.

[0276] For example, when the current deviation stroke in the character "待" is the first left-falling stroke and there is no deviation stroke the same as that of the writing user, then according to the basic left-falling stroke, the left-falling strokes corresponding to the first and second strokes in "待" can be extracted, and the "丿" of the first and second strokes are used as the selected strokes. At the same time, delete the current deviation stroke of the first left-falling stroke from the selected strokes, so as to obtain the remaining "丿" of the second stroke and use it as the similar stroke.

[0277] C02. Obtain the maximum value of the abscissa, the minimum value of the abscissa, the maximum value of the ordinate, and the minimum value of the ordinate of the dot matrix coordinates corresponding to each of the similar strokes.

[0278] It can be understood that obtaining the maximum value of the abscissa, the minimum value of the abscissa, the maximum value of the ordinate, and the minimum value of the ordinate of the dot matrix coordinates corresponding to the similar strokes is for subsequent comparison of the stroke standardization.

[0279] C03. According to the difference between the maximum value of the abscissa and the minimum value of the abscissa of the similar strokes, obtain the similar horizontal length value. According to the difference between the maximum value of the ordinate and the minimum value of the ordinate of the similar strokes, obtain the similar vertical length value. [[ID= 17]]

[0280] It can be understood that the similar horizontal length value is the difference between the maximum value of the abscissa and the minimum value of the abscissa of the similar strokes. For example, when the maximum value of the abscissa of the similar strokes is 10 and the minimum value of the abscissa is 1, the corresponding similar horizontal length value is 10 - 1 = 9. The similar vertical length value is the difference between the maximum value of the ordinate and the minimum value of the ordinate of the similar strokes. For example, when the maximum value of the ordinate of the similar strokes is 15 and the minimum value of the ordinate is 10, the corresponding similar vertical length value is 15 - 10 = 5.

[0281] C04. Obtain the maximum value of the abscissa, the minimum value of the abscissa, the maximum value of the ordinate, and the minimum value of the ordinate of the dot matrix coordinates corresponding to each of the current deviation strokes.

[0282] It is understandable that obtaining the maximum, minimum, maximum, and minimum values ​​of the horizontal and vertical coordinates of the corresponding dot matrix coordinates of the current deviation stroke is necessary to calculate the horizontal and vertical length values ​​of the current deviation stroke, so as to compare it with the corresponding strokes of the same type.

[0283] C05. The current horizontal length value is obtained based on the difference between the maximum and minimum horizontal coordinate values ​​of the current deviation stroke. The current vertical length value is obtained based on the difference between the maximum and minimum vertical coordinate values ​​of the current deviation stroke.

[0284] Understandably, the current horizontal length value is the difference between the maximum and minimum horizontal coordinates of the current deviation stroke. For example, if the maximum horizontal coordinate of the current deviation stroke is 9 and the minimum horizontal coordinate is 5, then the corresponding current horizontal length value is 9-5=4. The current vertical length value is the difference between the maximum and minimum vertical coordinates of the current deviation stroke. For example, if the maximum vertical coordinate of the current deviation stroke is 17 and the minimum vertical coordinate is 13, then the corresponding current vertical length value is 17-13=4.

[0285] C06. Based on the absolute value of the difference between the current horizontal length value and the horizontal length values ​​of the same type, a first approximate evaluation value is obtained.

[0286] Understandably, the first approximate evaluation value is the absolute value of the difference between the current horizontal length value and the horizontal length value of the same type. For example, when the horizontal length value of the same type is 9 and the current horizontal length value is 4, the first approximate evaluation value can be obtained as 9-4=5.

[0287] C07. A second approximate evaluation value is obtained based on the absolute value of the difference between the current longitudinal length value and the longitudinal length value of the same type.

[0288] Understandably, the second approximate evaluation value is the absolute value of the difference between the current vertical length value and the vertical length value of the same type. For example, when the vertical length value of the same type is 5 and the current vertical length value is 4, the second approximate evaluation value can be obtained as 5-4=1.

[0289] C08, based on the first approximate evaluation value and the second approximate evaluation value, calculate the evaluation coefficient corresponding to the same type of stroke.

[0290] It is understandable that the first approximate evaluation value and the second approximate evaluation value have corresponding first weight value and second weight value. The first weight value and the second weight value can be preset by humans. The evaluation coefficient is obtained by adding the value obtained by multiplying the first approximate evaluation value and the first weight value with the value obtained by multiplying the second approximate evaluation value and the second weight value.

[0291] Among them, the evaluation coefficient is the value of the similarity between the same type of strokes and the current deviated strokes.

[0292] Through the above implementation manners, the present invention can obtain the evaluation coefficients corresponding to the same type of strokes, so as to facilitate subsequent comparison of the same type of strokes, thereby determining approximate strokes, so as to obtain the corresponding writing same-type users.

[0293] C09. Determine the corresponding same-type strokes as approximate strokes according to the comparison information between the evaluation coefficient and the preset evaluation interval.

[0294] It can be understood that the preset evaluation interval is an interval of evaluation coefficients set in advance, which can be set artificially in advance.

[0295] It is not difficult to understand that when the evaluation coefficient is within the preset evaluation interval, it can be shown that the same-type strokes are similar to the current deviated strokes, and then the same-type strokes can be used as approximate strokes, so as to obtain the corresponding writing same-type users.

[0296] C10. Select the writing training users according to the approximate strokes to obtain the writing same-type users corresponding to the cross strokes.

[0297] It can be understood that after obtaining the approximate strokes, the writing training users can be selected according to the approximate strokes, so as to obtain the writing same-type users corresponding to the number of deviated strokes, and then obtain the corresponding training data for data reference, making the calculated customized writing times more accurate.

[0298] It should be noted that the above solutions are only applicable to characters with the same strokes. For example, "two", "wait", etc. For special characters without similar strokes, similar fonts can be selected, such as "big" and "too", etc.

[0299] See Figure 4 , which is a schematic structural diagram of a writing intelligent pen for assisted writing based on big data provided by an embodiment of the present invention, including:

[0300] An analysis module, configured to receive the test data of the writing user, obtain the to-be-trained characters in the test data, retrieve the training database to perform similarity analysis on the to-be-trained characters, and obtain the predicted training times from the deviation state to the standard state corresponding to each to-be-trained character.

[0301] An acquisition module, configured to acquire the historical training data of the writing user, and obtain adjustment data according to the predicted training times and the actual training times of the to-be-trained characters in the historical training data.

[0302] The adjustment module is used to adjust the predicted training times based on the adjustment data to obtain the planned training times for the corresponding character to be trained, and to obtain a customized writing page corresponding to the writing user based on the planned training times and the baseline writing times.

[0303] The generation module is used to customize and combine the characters to be trained in the customized writing page according to the reconstructed lexicon, and generate a combined writing page corresponding to the writing user.

[0304] See Figure 5 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present invention. The electronic device 50 includes: a processor 51, a memory 52, and a computer program; wherein...

[0305] The memory 52 is used to store the computer program, and the memory may also be flash memory. The computer program is, for example, an application program or functional module that implements the above method.

[0306] The processor 51 is configured to execute the computer program stored in the memory to implement the various steps performed by the device in the above method. For details, please refer to the relevant descriptions in the preceding method embodiments.

[0307] Alternatively, the memory 52 can be either standalone or integrated with the processor 51.

[0308] When the memory 52 is a device independent of the processor 51, the device may further include:

[0309] Bus 53 is used to connect the memory 52 and the processor 51.

[0310] The present invention also provides a readable storage medium storing a computer program, which, when executed by a processor, is used to implement the methods provided in the various embodiments described above.

[0311] The readable storage medium can be a computer storage medium or a communication medium. A communication medium includes any medium that facilitates the transfer of computer programs from one location to another. A computer storage medium can be any available medium accessible to a general-purpose or special-purpose computer. For example, a readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application-Specific Integrated Circuit (ASIC). Alternatively, the ASIC can be located in a user equipment. Of course, the processor and the readable storage medium can also exist as discrete components in a communication device. The readable storage medium can be a read-only memory (ROM), random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0312] The present invention also provides a program product including executable instructions stored in a readable storage medium. At least one processor of the device can read the executable instructions from the readable storage medium, and the at least one processor executes the executable instructions to cause the device to implement the methods provided in the various embodiments described above.

[0313] In the embodiments of the above-described device, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly manifested as execution by a hardware processor, or execution by a combination of hardware and software modules within the processor.

[0314] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A writing method for assisting writing based on big data, characterized by, include: Receive test data from writing users, obtain the characters to be trained in the test data, retrieve the training database to perform similarity analysis on the characters to be trained, and obtain the predicted number of training times from the deviation state to the standard state corresponding to each character to be trained. Obtain the historical training data of the writing user, and obtain adjustment data based on the predicted training times and actual training times of each character to be trained in the historical training data; Based on the adjustment data, the predicted training times are adjusted to obtain the planned training times for the corresponding character to be trained. Based on the planned training times and the baseline writing times, a customized writing page corresponding to the writing user is obtained. Based on the reconstructed lexicon, the characters to be trained in the customized writing page are customized and combined to generate a combined writing page corresponding to the writing user.

2. The method according to claim 1, characterized in that, The process of receiving test data from writing users, obtaining the characters to be trained from the test data, retrieving the training database to perform similarity analysis on the characters to be trained, and obtaining the predicted training count from the deviation state to the standard state corresponding to each character to be trained includes: Obtain the test data of the writing user, and retrieve the standard writing data corresponding to each test character in the test data; Based on the comparison information between the actual writing data of the test character and the corresponding standard writing data, the deviation writing data is obtained. The number of deviation writing data is counted to obtain the deviation writing quantity, and the number of the corresponding actual writing data is counted to obtain the actual writing quantity. The deviation ratio is obtained based on the ratio of the number of deviations written to the number of actual writings. When the deviation ratio is greater than the preset ratio, the corresponding test character is used as the character to be trained. Retrieve from the training database the writing training users corresponding to the character to be trained from the deviation state to the standard state, and the number of writing training times corresponding to the writing training users; Stroke analysis is performed on the actual writing data to obtain the current deviation strokes. Based on the current deviation strokes, similarity screening is performed on the writing training users to obtain writing similar users corresponding to each intersecting stroke. The number of writing training times corresponding to the writing type of users is determined as the number of writing type. Based on the number of writing type of each intersecting stroke and the corresponding initial weight value, the predicted number of training times corresponding to the character to be trained is obtained.

3. The method according to claim 2, characterized in that, The step of obtaining deviation writing data based on the comparison information between the actual writing data of the test character and the corresponding standard writing data includes: Obtain the writing pixel value corresponding to the actual writing data of the test character, and statistically analyze the dot matrix coordinates corresponding to the actual writing data based on the writing pixel value to obtain the writing coordinate set corresponding to the actual writing data. Retrieve the standard coordinate set corresponding to the corresponding standard writing data, determine the intersection of the standard coordinate set and the corresponding writing coordinate set, and obtain the standard coordinate set; Based on the difference between the writing coordinate set and the standard coordinate set, the deviation coordinate set is obtained. The number of lattice coordinates in the deviation coordinate set is obtained to obtain the number of writing deviations. The number of lattice coordinates in the standard coordinate set is obtained to obtain the number of writing standards. The writing deviation ratio is obtained by comparing the number of writing deviations with the number of writing standards. When the writing deviation ratio is determined to be greater than the preset writing ratio, the corresponding actual writing data is used as the deviation writing data.

4. The method according to claim 2, characterized in that, The step involves performing stroke analysis on the corresponding actual writing data to obtain the current deviation strokes, and then filtering the writing training users based on the current deviation strokes to obtain writing users of similar type corresponding to each intersecting stroke, including: Obtain the set of stroke coordinates corresponding to the actual writing data, and retrieve the standard set of strokes corresponding to the standard writing data, as well as the number of dot matrix coordinates in the standard set of strokes to obtain the number of standard dots. The stroke specification set is obtained by the intersection of the stroke standard set and the stroke coordinate set; Based on the difference between the set of stroke coordinates and the corresponding set of stroke specifications, a set of stroke deviations is obtained, and the number of dot matrix coordinates in the set of stroke deviations is obtained to obtain the number of deviation dots. The number of stroke errors corresponding to the corresponding writing stroke is calculated based on the standard number of dots, the preset number of dots, and the preset number of errors. When the number of deviation dots is determined to be greater than the number of strokes tolerable, the corresponding written stroke is regarded as a misaligned stroke. The number of misaligned strokes is determined, and the stroke deviation ratio corresponding to the misaligned strokes is obtained based on the ratio of the number of misaligned strokes to the corresponding actual number of written strokes. When the stroke deviation ratio is determined to be greater than the preset stroke ratio, the corresponding misaligned stroke is taken as the deviation stroke. Obtain the deviation strokes of the writing training users as historical deviation strokes, and determine the corresponding writing training users as initial screening training users based on the intersection relationship between the historical deviation strokes and the current deviation strokes. The strokes that intersect with the current deviation strokes of the initial screening training users are obtained as intersecting strokes, and the number of intersecting strokes is obtained as the number of deviation strokes. The initial screening training users are sorted in descending order according to the number of deviation strokes to obtain the selection user sequence. Based on a preset selection quantity, the initial screening training users in the selected user sequence are selected sequentially to obtain the writing type users corresponding to each of the intersecting strokes.

5. The method according to claim 4, characterized in that, The step of obtaining the historical training data of the writing user, and obtaining adjustment data based on the predicted training times and actual training times of each character to be trained in the historical training data, includes: Obtain the predicted training times and actual training times for each character to be trained from the historical training data of the writing user; When it is determined that the actual number of training iterations is greater than the predicted number of training iterations, a positive training value is determined based on the difference between the actual number of training iterations and the predicted number of training iterations. The initial weight value corresponding to the largest number of deviation strokes is obtained as the positive training weight. The positive training weight is increased and adjusted according to the positive training value to obtain the adjusted positive training weight. The weight adjustment value is obtained based on the difference between the adjusted positive training weight and the positive training weight. The initial weight values ​​are sorted in ascending order based on the number of deviation strokes to obtain a weight adjustment sequence; The first initial weight value in the weight adjustment sequence is selected as the reverse training weight, and the adjusted reverse training weight is obtained based on the difference between the reverse training weight and the weight adjustment value. When it is determined that the adjusted reverse training weight is less than 0, the first initial weight value in the weight adjustment sequence is deleted, the repeated adjustment value is obtained according to the absolute value of the adjusted reverse training weight, and the corresponding adjusted reverse training weight is set to 0. Use the repeated adjustment value as the current weight adjustment value, and repeat the above steps of using the repeated adjustment value as the current weight adjustment value until the adjusted reverse training weight is greater than or equal to 0, thus obtaining multiple adjusted reverse training weights. Adjusted data is obtained based on the adjusted forward training weights and the adjusted backward training weights.

6. The method according to claim 4, characterized in that, The step of obtaining the historical training data of the writing user, and obtaining adjustment data based on the predicted training times and actual training times of each character to be trained in the historical training data, includes: When it is determined that the actual number of training iterations is less than the predicted number of training iterations, a reverse training value is determined based on the difference between the predicted number of training iterations and the actual number of training iterations. The initial weight value corresponding to the largest number of deviation strokes is obtained as the reverse training weight. The reverse training weight is adjusted by decreasing the reverse training value to obtain the adjusted reverse training weight. The weight adjustment value is obtained based on the difference between the reverse training weight and the adjusted reverse training weight; The initial weight value corresponding to the minimum number of deviation strokes is obtained as the positive training weight. The adjusted positive training weight is obtained based on the weight adjustment value and the sum of the positive training weight. Adjusted data is obtained based on the adjusted forward training weights and the adjusted backward training weights.

7. The method according to claim 1, characterized in that, The step of obtaining a customized writing page corresponding to the writing user based on the planned training count and the baseline writing count includes: The total number of training iterations corresponding to the characters to be trained is obtained by counting the planned training iterations. Based on the ratio of each planned training iteration to the total number of training iterations, the proportion coefficient corresponding to the characters to be trained is obtained. The customized number of writings corresponding to the character to be trained is obtained by multiplying the proportion coefficient and the baseline number of writings. The corresponding characters to be trained are combined according to the number of customized writing attempts to obtain a customized writing page.

8. The method according to claim 1, characterized in that, The step of customizing and combining the characters to be trained in the customized writing page according to the reconstructed lexicon to generate a combined writing page corresponding to the writing user includes: Based on the preset word groups in the reconstructed lexicon, the characters to be trained in the customized writing page are combined to obtain multiple customized word groups, each of which includes multiple characters to be trained. The number of characters to be trained in the customized word group is obtained to get the number of combinations. The customized word group is sorted in descending order according to the number of combinations to get the customized sequence. Select the first customized word group in the customized sequence as the current customized word group, and take the character to be trained in the current customized word group as the current character to be trained. Obtain the customized writing count corresponding to the current character to be trained, select the current character to be trained with the smallest customized writing count as the deletion and update character, and use the remaining current characters to be trained as relay update characters; The number of times the customized writing of the deleted and updated characters is used as the update count. The current customized word group is generated based on the update count, and the number of times the customized writing of the relay updated characters is updated based on the update count. Delete the custom word groups that contain the deleted and updated characters in the custom sequence, and repeat the above steps of deleting the custom word groups that contain the deleted and updated characters in the custom sequence until the custom sequence is empty, to obtain multiple current custom word groups, and use the remaining characters to be trained as random characters to be trained; The combined writing page is obtained by combining the currently customized phrases and the random characters to be trained.

9. The method of claim 5, wherein, Also includes: Obtain the number of initial screening users for initial screening. When the number of initial screening users is equal to 0, determine the basic stroke unit corresponding to the current deviation stroke. Extract the corresponding writing strokes in the characters to be trained as selection strokes based on the basic stroke unit. Delete and update the selection strokes based on the current deviation strokes to obtain strokes of the same type. Obtain the maximum, minimum, maximum, and minimum values ​​of the horizontal coordinates, vertical coordinates, and ordinates of the corresponding dot matrix coordinates for each stroke of the same type. The horizontal length value of the same type of stroke is obtained by the difference between the maximum and minimum horizontal coordinate values ​​of the same type of stroke; the vertical length value of the same type of stroke is obtained by the difference between the maximum and minimum vertical coordinate values ​​of the same type of stroke. Obtain the maximum, minimum, maximum, and minimum values ​​of the horizontal and vertical coordinates of the corresponding dot matrix coordinates for each of the current deviation strokes; The current horizontal length value is obtained based on the difference between the maximum and minimum horizontal coordinate values ​​of the current deviation stroke, and the current vertical length value is obtained based on the difference between the maximum and minimum vertical coordinate values ​​of the current deviation stroke. A first approximate evaluation value is obtained based on the absolute value of the difference between the current horizontal length value and the horizontal length values ​​of the same type; A second approximate evaluation value is obtained based on the absolute value of the difference between the current longitudinal length value and the longitudinal length value of the same type; The evaluation coefficient corresponding to the same type of stroke is calculated based on the first approximate evaluation value and the second approximate evaluation value. Based on the comparison information between the evaluation coefficient and the preset evaluation interval, the corresponding strokes of the same type are determined as approximate strokes; Based on the approximate strokes, the writing training users are selected to obtain writing users of the same type as the intersecting strokes. 10.A writing smart pen for assisting writing based on big data, characterized by, include: The analysis module is used to receive test data from the writing user, obtain the characters to be trained in the test data, retrieve the training database to perform similarity analysis on the characters to be trained, and obtain the predicted number of training times from the deviation state to the standard state corresponding to each character to be trained. The acquisition module is used to acquire the historical training data of the writing user and obtain adjustment data based on the predicted training times and actual training times of each character to be trained in the historical training data. The adjustment module is used to adjust the predicted training times based on the adjustment data to obtain the planned training times for the corresponding character to be trained, and to obtain a customized writing page corresponding to the writing user based on the planned training times and the baseline writing times. The generation module is used to customize and combine the characters to be trained in the customized writing page according to the reconstructed lexicon, and generate a combined writing page corresponding to the writing user.

Images